This research proposal is describes three projects that are concerned with the structural knowledge and perceptual process that mediate word recognition, perceptual learning of speech, aNd perception of talker characteristics. The first project is concerned with understanding the role of structural knowledge in word perception. Two experiments in this project investigate how structural properties of words are used to segment an utterance into word-length patterns and how these properties may be used to facilitate the process of word recognition. A third experiment is directed at investigating whether the patterns of words in different lexical classes (e.g., open and closed class words) can differentially facilitate recognition of the words in these classes and whether structural differences allow listeners to make word class judgments during recognition, rather than as a result of post-recognition lexical access. The fourth experiment examines how lexical familiarity and frequency recognition of spoken words and the processing of their constituent elements. The second project is concerned with the processes and knowledge that support perceptual learning of speech. The first experiment will investigate how a listener's knowledge of the structural properties of speech may be used to learn to overcome distortions in the phonetic structure of speech. The second experiment is concerned with the role of the organization of the mental lexicon in learning new words. The final experiment will investigate how context-conditioned variability of the pattern structure of a signal affects the perceptual learning and recognition of the information encoded in the signal. Although this experiment will use nonspeech acoustic patterns, the results may suggest how the perceptual system could use the acoustic consequences of coarticulation to facilitate learning and recognition of the acoustic-phonetic structure of speech. The third projects directed at understanding how perception of the characteristics of a talker's voice can affect recognition of the lexical and phonetic content of the talker's speech. The first experiment will investigate possible interactions between the attentional demands of recognizing the phonetic structure of speech and recognizing a particular talker's voice. The second experiment is concerned with the acoustic properties of speech that are used by a listener to focus on a particular talker's voice among other voices and how these properties affect recognition of the information conveyed by that voice. The final experimental will investigate how memory for a particular talker's voice affects perceptual normalization of talker difference to facilitate recognition of the segments in speech. The findings of these three projects haVe broad implications for the diagnosis and treatment of disorders and diseases that affect the perception and comprehension of spoken language.